Everyone in marketing has felt the shift. AI Overviews show up where blue links used to be, ChatGPT and Perplexity answer questions that used to drive Google clicks, and "what does our website even do anymore" is a more common boardroom question than anyone expected in 2024.
What has been harder to come by is honest numbers. The vendor pitches are full of round figures. The trade press alternates between "search is dead" and "everything is fine." Neither is true. This post collects what is actually documented in the industry research, layers in a few well-reported public case studies, and gives realistic timeframes for what each layer of the stack — SEO, AEO, GEO — can move and how fast.
One note before the numbers. Methodology matters here. Where we cite a statistic, we name the firm that published it. Where we describe a case study, we either point to public reporting or label it explicitly as a composite. Where the data is still emerging — which is most of GEO — we hedge accordingly. The goal is a defensible picture, not a tidy one.
The Macro Picture: Traffic Is Redistributing, Not Disappearing
Total search demand has not collapsed. What changed is where the demand resolves. Google still dominates query volume, but a meaningful and growing share of question-shaped queries now resolve inside AI surfaces — Google AI Overviews, Perplexity, ChatGPT, Claude — without a click to a publisher.
Google's overall search volume has continued to grow in absolute terms since 2023, according to its own quarterly disclosures and third-party panel data from firms like Similarweb. What dropped is the click-through rate to publisher sites for specific query categories — primarily informational and how-to queries — as AI Overviews intercepted the answer.
Three macro shifts to anchor the rest of the numbers:
BrightEdge's tracking, widely cited through 2024 and 2025, reported AI Overview prevalence varying dramatically by vertical — with healthcare, B2B technology, education, and finance categories sometimes exceeding 60% of queries while ecommerce transactional queries stayed under 10%. The headline number depends entirely on which industry you are in.
SEO: What Is Still Working, What Is Eroding
SEO is not dead. For most sites it remains the largest organic traffic channel by a wide margin. But the unit economics of traditional SEO have measurably shifted, and the shift accelerated after May 2024.
The Click-Through Rate Picture
The pre-AI-Overview baseline, established by long-running studies from Advanced Web Ranking, Backlinko, and Sistrix between 2019 and 2023, put average organic CTR at roughly:
| POSITION | HISTORICAL AVG CTR | SOURCE PATTERN |
|---|---|---|
| 1 | ~27–32% | Backlinko 2023, AWR 2022, Sistrix 2020 |
| 2 | ~15–18% | Backlinko 2023, AWR 2022 |
| 3 | ~10–11% | Backlinko 2023, AWR 2022 |
| 4–10 | ~2–8% each | Industry consensus pre-2024 |
Multiple SEO firms — including Ahrefs, Semrush, and Search Engine Land's editorial research — published findings through 2024 and 2025 showing that when an AI Overview appears, click-through rates on the underlying organic results typically fall by roughly 30 to 70 percent, with the steepest declines on pure-information queries. Branded queries and transactional queries showed much smaller effects.
These CTR declines are not uniform. Methodologies vary across the firms reporting them, and "the queries that show an AI Overview" is itself a moving target as Google adjusts which categories surface the feature. Use the 30–70% range as a directional planning number, not a precise forecast for your specific site.
Public Case Studies
Stack Overflow. The most widely-reported organic traffic decline in the post-ChatGPT era. Independent measurement firms including Similarweb tracked Stack Overflow visits as falling sharply through 2023 and 2024 — the company itself acknowledged the dynamic publicly, and observers including The Register and tech press attributed much of the decline to developers asking ChatGPT instead. Order of magnitude: roughly 50% lower traffic than 2022 peaks by late 2024 in widely-cited third-party measurements. Timeframe from disruption start to measurable impact: about 12 to 18 months.
Chegg. The education services company publicly attributed a portion of its 2023 revenue decline to ChatGPT in its earnings disclosures. The stock fell roughly 50% on a single day in May 2023 after the CEO discussed the impact. Subscriber growth slowed measurably through 2023 and 2024. This was one of the first publicly-traded admissions that generative AI was directly displacing a paid product.
The publisher cohort. Through 2024 and 2025, multiple major publishers — including reporting from Press Gazette, Digiday, and The Atlantic's own coverage of its traffic data — described measurable Google referral traffic declines following AI Overview rollout. The specific drops varied widely by publisher and category, but the directional consensus across the trade press was a meaningful and persistent reduction in traditional Google organic for informational verticals.
AEO: The Citation Economy
AEO investment produces faster, more measurable results than either SEO or GEO. The lift is not in clicks — it is in zero-click brand visibility plus a smaller but high-intent click stream from cited links. Pages with proper schema get cited; pages without it largely do not.
The Numbers
Featured snippets — the long-running ancestor of AI citations — have been studied for years. Ahrefs and other SEO firms have reported featured snippet CTR for the snippet itself typically falling in the 8% to 35% range, with wide variation by query type. A page promoted into a featured snippet often saw measurable traffic lift even when the snippet answered the question, because the snippet displayed the source's name and domain prominently.
For modern AI citations, the picture is messier but the directional findings are consistent across multiple studies:
- Pages with FAQ schema are cited by Perplexity and Google AI Overviews at meaningfully higher rates than pages without it — findings consistent across reporting from Search Engine Land, SE Ranking, and independent practitioner posts through 2024 and 2025.
- The click-through rate from an AI citation is lower than from a traditional blue-link result — by some estimates only a third to a half — but the visitors who do click arrive with higher intent and longer session durations, according to multiple analytics-vendor reports.
- Voice-assistant answer attribution remains low but is growing; Speakable schema adoption is rare enough that early movers in supported verticals have outsized share of voice.
Timeframe to Impact
| AEO MOVE | TIME TO MEASURABLE EFFECT |
|---|---|
| Add FAQ schema, test for Perplexity citation | 2–14 days from recrawl |
| Add 40–60 word answer blocks under question headings | 2–6 weeks for featured snippet capture |
| Strengthen author / organization schema | 4–12 weeks for AI Overview citation lift |
| HowTo schema on tutorial pages | 2–8 weeks for rich result eligibility |
| Speakable markup for voice surfaces | 4–8 weeks for voice attribution increase |
Composite Case Pattern
The pattern we have seen repeatedly when developer-tooling and SaaS documentation sites add a full AEO layer — FAQ schema on documentation pages, question-format headings with concise answer blocks, strengthened organization schema — is a measurable lift in AI citation rate within a four-to-eight-week window, with Perplexity moving fastest and Google AI Overviews moving slowest. Order of magnitude in well-instrumented examples: citation rate on the prompt suite roughly doubling over an eight-week period, off a low base. The traffic implications depend entirely on the prompt-to-conversion path, but for high-intent documentation queries the cited traffic often converts at 2x or more the rate of generic organic.
This is a composite description, not a single named customer's data. The directional pattern is consistent enough across multiple practitioner reports that we are comfortable describing it.
GEO: The Hardest to Quantify, Maybe the Most Valuable
GEO results compound slower than AEO and are harder to measure cleanly. But generative engines now influence a measurable share of B2B and consumer purchase research, and brands the models recommend in their generated answers earn mindshare that traditional analytics largely cannot see.
The Volume Picture
ChatGPT's weekly active user count was reported by OpenAI as crossing 200 million in late 2024, with continued growth disclosed in subsequent public statements. Perplexity's query volume grew an order of magnitude over 2024 by its own and third-party reporting. Claude usage growth, while less publicly disclosed, has tracked broadly similar adoption curves in enterprise and developer segments.
Survey research from firms including Bain, McKinsey, and various B2B-focused research practices through 2024 and 2025 consistently found that a meaningful and growing minority of B2B buyers — somewhere in the 20–40% range, depending on the survey and vertical — reported using generative AI tools as part of their software discovery and shortlisting research. The signal is real even if the precise number depends on who is asking.
Timeframe to Impact
| GEO MOVE | TIME TO MEASURABLE EFFECT |
|---|---|
| Entity consistency cleanup across LinkedIn / Crunchbase / your site | 2–4 weeks (branded-prompt accuracy lift) |
| Single landed inclusion in a category listicle | 2–6 weeks (mention rate lift) |
| Trade publication feature | 4–12 weeks (sentiment and citation lift) |
| Original research with citable data | 3–9 months (compounding third-party citations) |
| Wikipedia entry (where eligible) | 3–6 months (durable entity authority lift) |
The Attribution Reality
The honest GEO numbers problem is that most of the value cannot be measured by traditional analytics. A prospect who asks ChatGPT for software recommendations, hears your brand, then types your domain into the browser bar will show up in your analytics as direct traffic — indistinguishable from a returning user. Some firms now estimate that "AI-attributed" direct traffic accounts for a measurable share of new B2B inbound in 2025 and 2026, but the methodologies are early and the numbers are best read as directional, not precise.
Composite Trajectory: Three Hypothetical Sites Over 12 Months
To make the stakes concrete, here is a composite picture of three otherwise-identical sites pursuing different strategies over a 12-month window. Numbers are illustrative — derived from the ranges documented above — not measured from a specific real site.
| STRATEGY | SEO TRAFFIC | AI CITATION RATE | BRAND MENTIONS IN AI |
|---|---|---|---|
| Site A: SEO only — does nothing new | Down 20–40% | Effectively zero | Effectively zero |
| Site B: SEO + AEO — adds full schema layer | Down 5–15%, partially offset | 2–4x baseline | Modest lift on branded prompts |
| Site C: SEO + AEO + GEO — full stack | Flat to up 10% | 3–6x baseline | Materially higher across all prompt families |
The trajectories diverge most sharply on the channels traditional analytics struggle to measure. Site A's dashboards look bad. Site B's dashboards look okay. Site C's dashboards look fine — but the actual share-of-voice gap between Site C and Site A is substantially larger than the dashboards suggest, because so much of the GEO value lives in zero-click visibility.
The Cost of Inaction
The compounding pattern is the part most sites underestimate. SEO erosion does not arrive all at once. AI Overview rollout has been incremental, category by category. ChatGPT and Perplexity recommendation patterns harden gradually as the models update and as the citation graph thickens. A six-month delay in adding AEO and starting GEO measurement does not just delay your results by six months — it gives the sites that did move six months of compounded entity advantage that gets progressively harder to overtake.
That is the part that matters most for the budget conversation. The question is not "do these new layers work" — the documented research increasingly says they do, on the timeframes outlined above. The question is whether the cost of moving now is lower than the cost of letting the gap widen for another two quarters.
Overt Ops measures all three layers on your site continuously — SEO health, AEO citation readiness, and GEO presence across the four major engines — and surfaces the moves most likely to compound based on what is already documented to work in your category. One tool, three layers, real numbers. Coming soon from Area 51 Software.
Methodology & Sources
The statistics in this post are drawn from publicly available industry research published through 2024 and 2025 by SEO and analytics firms including BrightEdge, Ahrefs, Semrush, Advanced Web Ranking, Backlinko, Sistrix, Similarweb, and Search Engine Land's editorial research; from publicly-disclosed company filings and statements (Chegg, Google, OpenAI); from trade press reporting (Press Gazette, Digiday, The Atlantic, The Register); and from survey research by Bain, McKinsey, and B2B-focused research practices. Where we cite a range — "30 to 70 percent CTR decline" — the range reflects genuine variation across studies, not imprecision in any single source. Where we describe a case pattern as composite or illustrative, we mean it: the directional shape is consistent with multiple practitioner reports but is not a single named customer's data. The AI search landscape is moving fast enough that any number in this post is worth re-checking against current research before it shapes a budget decision.
What to Do With These Numbers
Three honest takeaways:
The downside of doing nothing is documented. Sites in informational and how-to verticals that have not adapted are losing measurable organic traffic share to AI Overviews, on timeframes measured in months. The Stack Overflow and Chegg cases are the visible tip of a much broader, less public pattern.
The upside of moving on AEO is the fastest and most measurable lever available. Schema work, answer-block restructuring, and entity strengthening produce measurable changes inside an eight-week window for most sites that move with discipline. There is no version of AEO investment that hurts SEO.
The GEO investment is the longest-horizon and the one most prone to over-confident claims. Build the measurement system, fix entity consistency, run a deliberate citation campaign — and revisit quarterly. The numbers will get better as the discipline matures.
The strategy that consistently wins is the same one we ended the last post with: SEO is the foundation, AEO is the bridge, GEO is the new layer. Build them in order. Measure them honestly. Move now rather than later, because the gap compounds in only one direction.
OVERT OPS — COMING SOON
Continuous SEO + AEO + GEO measurement across the four major engines. Documented benchmarks, prioritized fixes, and honest reporting on what is moving and what is not.
Coming Soon